GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Sizing a Mitsubishi Ducted Minisplit

dtkrie | Posted in Mechanicals on

We are replacing a 20+ year old 80% AFUE furnace and 3T A/C with a new ducted mini-split. I’ve interviewed 4-5 different contractors; the local Bryant rep was by far the most interested in asking questions and doing some homework to get us the proper system. This contractor also happens to be a Mitsubishi diamond contractor, and I think we’ve settled on a Mitsubishi M series HP w/ hyper heat coupled with an SVZ fan coil.

My question is related to sizing. Per our attached Manual J, we are looking at  a 2.5T or 3T unit. The 2.5T  qualifies for $600 in local rebates/tax credits; the 3T has slightly lower efficiency ratings and does not qualify. As I understand, the EER/SEER/HSPF ratings are taken at  max btu output, so partial-load efficiency of the 3T unit should be similar to the 2.5T. Is this accurate?

The 3T is way oversized for cooling, but it allows the heat pump to carry our full heating load right down to our design temp. The 2.5T seems to better fit our partial loads in the swing seasons, but it’s latent cooling capacity barely meets the manual J load, and we’ll have to supplement with electric backup heat once temps drop below 5F. Our contractor recommends the 3T due to the humidity removal  & higher heating output.

Suggestions on which size is best?

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Expert Member
    Akos | | #1

    Take the $600 rebate plus the extra bit saved from the smaller unit and get a decent ERV. Heat recovery would shave about 1/2 ton of your heat load.

    Even if that feels too tight for comfort for you, running a resistance strip as extra boost for those few days bellow design is noise on your yearly power consumption.

    Humidity removal is a function of CFM/ton, you always turn down the external static pressure setting on the air handler to reduce the flow a bit if it becomes an issue in the summer.

    Since your cooling loads are so low, the 3 ton would actually be much worse for humidity removal as the SHR tends to increase on most units at low modulation.

  2. paul_wiedefeld | | #2

    7 occupants is correct? Seems like heat gain in winter from them would cover the slight difference in outputs. If they’re going to install the 8kw backup, I think the 2.5 is better.

    1. dtkrie | | #3

      Yes, 7 occupants is correct for our immediate family.

  3. dtkrie | | #4

    Looking at the min. cooling loads at 95F, there actually isn't much difference between the 2.5T & 3T (13.4k v. 14.2k). Looks like we'd have to step down to the 2T M-series (8,800 min. output @ 95F) to really align with our cooling load, but 2T is nowhere close to satisfying our heating demand.

    Side note, we also have a HPWH which draws an additional 4200 btu/hr from our conditioned basement while it is running (typically between 7 PM - 12 PM) Obviously this makes the cooling demand discrepancy even greater, but I was actually more concerned about the extra 4k heating load which is not currently included in our manual J. I figured the added heat from the occupant load would somewhat offset the heat loss from the HPWH, but perhaps that was naive?

  4. PBP1 | | #5

    Seems like the total heating load estimate may be too high. The floor related load (almost 8kBTU/hr) seems high, especially given that part of the first floor is over the basement (plans do not appear to show stairs from basement to first floor?). I'm in Montana with 1,350 first floor (mostly on piers, i.e., exposed) and 750 second floor (mostly above a 2 car garage), with 6 6' wide sliders (a lot of glass) on first floor, the heat load is 28KBTU/hr satisfied with a Mitsubishi ASHP rated at 28kBTU @ 47F with three zones ducted (15K, 12K, 9K). My first floor will see -10F as it is elevated; whereas, your basement floors are in contact with the ground and will not see such low temperatures. As such, heat loss via your ground-contact floors will be limited/bufferred as temperature drops and floors above the basement are going to be even more bufferred. Thus total heating load of 34KBTU/hr seems too high for Lincoln, NE (where my nephew just graduated), but I might be missing something? Maybe 2T will actually satisfy your heating load, again, less than 2.5T for 2,100 sq ft (ACH50 < 2.5) with too much glass and exposed floor in Montana is fine, even below 0 F. (I mention stairway because heat is going to rise if it is open, my second floor has a separate entrance, no stack effect). My energy guy/HERS rater's estimate on my new build was within 5% of actual, he worked with HVAC guys (long established relationship), who have not had problems with ASHP in Montana winters (they know how to size).

    Some info: covers basements and slabs.

    Basements: "The temperature of an unheated below-grade basement is between the temperatures of the rooms above and the ground temperature. Heat losses from the water heater and the space heater located in the basement usually keep the air near the basement ceiling sufficiently warm. Heat losses from the rooms above to the basement can be neglected in such cases. This will not be the case, however, if the basement has windows."

    Slabs: "Many residential and commercial buildings do not have a basement, and the floor sits directly on the ground at or slightly above the ground level. Research indicates that heat loss from such floors is mostly through the perimeter to the outside air rather than through the floor into the ground, as shown in Fig. 39. Therefore, total heat loss from a concrete slab floor is proportional to the perimeter of the slab instead of the area of the floor"

    1. dtkrie | | #6

      The floor plan in the manual J is a little deceiving, as the software is not conducive to drawing a multi-level house. See attached floor plan for a more accurate depiction of our layout, including the two 1/2 height open stairways.

      I did compare our manual J using monthly gas use/HDD and by clocking the runtime on our current furnace. Based on furnace runtime, peak load at 0F is around 40k Btu/hr, but that assumes our 21-year-old 80% furnace is still actually performing at 80% efficiency. When I ran calcs using HDD and monthly gas use, the peak load ranged from 25k-29k. I'm not entirely convinced my HDD calcs are accurate, as I struggled to find a design temp which I felt was accurate.

      The majority of our slab heat loss is due to the living room and laundry room, which are slab-on-grade with no slab edge insulation. I hope to eventually dig down and install 3" EPS 12" below grade on the east elevation which is still accessible. This should reduce the heating load by about 3k; however, that still leaves overall heat load north of 30k Btu (assuming current manual J is accurate).

      What's your wall assembly and roof insulation on your Montana residence? I had to remodel our basement a few years ago, so our basement walls are R-39 (above-grade stick framed) & R-19 (concrete stem wall w/ interior framing); however, the rest of our walls are just R13 batt insulation with R1.2 blackjack sheathing on the exterior.

      1. PBP1 | | #7

        Thanks for sharing details, I’ll take a closer look. My floors are R38, walls R25 and ceiling R46, all flash and batt. I also have a 12’x12’x3’ concrete crawl with concrete floor, abuts garage pad, all insulated at perimeter. Some trades thought the heat pump wouldn’t be enough, trying to up sell back up heat. But HVAC and HERS guys were adamant, and correct. The amount of glass is a bit excessive, which adds many headers and hence lower actual R. Great room is 20’x30’ with 5 sliders and clerestory windows, lid rising from 9’ to 13.5’ with 5 6”x12” exposed beams that extend to the exterior and hence thermal bridging. Not the greenest, designed before I discovered this website. If did over, would have added continuous exterior insulation.

      2. Expert Member
        Akos | | #9

        Fuel based load calculations are much more accurate than any Man J as it is for the actual building as is without needing any assumptions.

        If your fuel load calc comes in at 25k to 29k, you can shave 3k off that with the slab edge insulation, that puts you firmly into the range of the 2.5 ton unit even without electric backup.

        Keep in mind that the difference between the 2.5 ton and the 3 ton unit is the equivalent of one plug in electric space heater, if it is really needed on those polar vortex days.

        1. dtkrie | | #10


          I was hoping to rely on HDD and furnace runtime for the reasons you noted above, but the HDD peak load varied so much (up to 21%) from month to month that I have little confidence in its accuracy. In the attached screenshot, I actually ran HDD calcs for multiple base points ranging from 51F to 63F. I'm guessing our actual base temp is somewhere in the 50's based on when I generally notice the need to turn on our furnace, but that is just my WAG.

          Furnace runtime (22 min. ON / 23 min OFF @ 3F outdoor temp) would suggest our peak heating load is closer to 40k Btu/hr, but that assumes the furnace is still 80% efficient. If the AFUE has actually dropped to ~70%, then our peak load would be ~35,000 btu/hr. I was measuring furnace runtime with a furnace that has a known issue with the gas valve, so short of asking our contractor to calculate actual btu using supply air temp and actual CFM's, I cannot verify the accuracy of the runtime-based load.

          1. PBP1 | | #17

            Any time you have a flame, losses can be substantial. A methane (natural gas) flame can be around 3500F. Capturing that heat energy can be difficult.

            From DOE - High efficiency furnace:
            Condensing flue gases in a second heat exchanger for extra efficiency
            Sealed combustion
            90% to 98.5% AFUE.

            Some say: "If you add these off-cycle losses to the steady state losses you end up with the seasonal efficiency. Seasonal efficiencies for conventional gas and oil furnaces are about sixty to sixty-five percent."

            Perhaps data for a high efficiency furnace with a supply temperature the same as a heat pump and similar steady-state type of operation can be a proxy for heat pump usage? Otherwise, I think gas/oil data are likely to be too high. Maybe 0.9 or so would be appropriate as an adjustment factor?

            A gas furnace for residential typically cycles such that a steady-state is not reached and such that the supply temperature tends to be much higher than that of a heat pump (greater delta T, which means greater heat transfer and generally greater losses).

            From AGA: "Natural gas heat feels warmer than heat produced by an electric heat pump. Natural gas heat is delivered from forced-air systems at temperatures ranging from 120-140 degrees Fahrenheit. In contrast, the air from an electric heat pump is typically delivered at 85-95 degrees Fahrenheit warm enough to heat a room, but cooler than the average human skin temperature of 98.6 degrees Fahrenheit."

            Heat transfer depends on driving force (delta T) thus with a supply temperature of 120-140F, there will likely be greater losses in the entire system than with 85-95F. Also, for a gas furnace, a distribution network may cool between cycles, which will increase the delta T between the supply air and the distribution network. The distribution network of a gas furnace can be subject to reheating for each cycle, where delta T is higher than for a heat pump, meaning potential for greater losses (rate of heat transfer) based on delta T alone.

            In general, things work best in steady-state. And, with the low supply temperature of a heat pump, something closer to steady-state operation tends to be favored.

            Anyone with details of ANSI/ASHRAE 103-2017 Method of Testing for Annual Fuel Utilization Efficiency of Residential Central Furnaces and Boilers? Just wondering how AFUE is actually calculated.

          2. Expert Member
            Akos | | #18

            Furnace runtime is a very noisy signal, at best gives you a ballpark range for your heat loss. Fuel use based calculation gets you closer, but you still have to estimate your furnace efficiency.

            I doubt your balance point is in the 50s unless you have something closer to a passive house. I'm guessing the reason for the late start to your heating season is because of solar gain. Solar gain is great in the fall but doesn't help much on those stormy nights when you get close to design temperature and there is no sun for a week, so a higher balance point temperature is probably more realistic.

            Another way also to look at your options is how much resistance heat runtime the price difference between the two units gets you. There is a 5000BTU output difference between the 2.5T and the 3T, which is ~1.5kW. The $900 of $0.15 of electricity buys you almost 170days of runtime. I doubt you'll see that within the lifetime of the equipment.

            If in doubt, have them install the small heater kit (5kW) which would boost the 5F output of the 2.5T from 32000BTU to 49000BTU. Plenty to cover those polar vortex days.

            P.S. There are other cold climate units you can also look at such as:


          3. dtkrie | | #19

            As much as I would like to entertain other manufacturers, I'm pretty set on Mitsubishi due to the contractor I am using. Of all the contractors I interviewed, only one was even remotely interested/competent in sizing the new system and diagnosing other potential problems with the existing ductwork. This contractor sells Bryant and Mitsubishi, so I'm inclined to stick with Mitsubishi.

  5. conwaynh85 | | #8

    Ive done two houses with ducted heat pumps and love them!
    I slightly undersized the first one and then added a completely seperate small head unit and seperate outdoor unit to supplement for the extreme weather.
    The second one was "right sized" with electric baseboard backup. With below zero temps, they both performed well.
    I prefer having "undersized ducted" two systems and using the head unit when needed. But if you dont want to pay the upfront cost, i would still go smaller with some supplement for extremes. Good luck!
    Im zone 6 inland

    1. dtkrie | | #11

      Thanks for the input. Between our utility rebates and the slightly lower equipment cost, I believe the 2.5T will be ~$900 less expensive than the 3T, which doesn't hurt the argument either. The 3T (16 SEER) actually has a slightly higher overall efficiency than the 2.5T (15 SEER); however, the 3T doesn't meet the minimum EER requirement for the federal tax credit or our local utility rebate.

      1. paul_wiedefeld | | #12

        Another example of why heat pump rebates are awful!

        1. dtkrie | | #13

          In defense of the rebate system, I assume the rebates are tied to EER because EER more closely resembles the efficiency when the unit is at max capacity (95F). Max capacity is ultimately what the electric utility companies want to minimize; if they can flatten their demand curve, they avoid the high expense of adding capacity to the grid. While it is unfortunate that this system sometimes rewards units that are less efficient overall (i.e. lower SEER), I can understand why utility companies tie the rebate to the EER rating as well as the SEER.

          1. paul_wiedefeld | | #14

            Max capacity during the summer is EER, yes. During the winter, you have a 8kw backup heater plus a 4kw heat pump for the 2.5 ton. Worried about 2.1 vs. 2.5kw during the summer seems silly.

  6. dtkrie | | #15

    In the US, the majority of buildings still rely on fossil fuels for their heating source, so the peak electrical demand occurs in the summer during cooling season. As more buildings convert to electric heating, eventually we'll hit a tipping point where peak load shifts to the winter, but I have no idea when that will occur.

    1. paul_wiedefeld | | #16

      True. Still a boneheaded system. If peak cooling demand is the issue, charge a $/kw demand charge like industrial users pay and actually deal with the problem. Instead (using 12 EER as an example cutoff), a 11.9 EER vs. 12 EER 2 ton unit saves 17 watts which an example $600 rebate pays almost $36,000/kw for. BUT the 13 EER system gets $0 additional rebate despite saving 10x the peak watts. Naturally, manufacturers suddenly predominantly make 12 EER systems. It also screws with homeowner decisions needlessly.

  7. dtkrie | | #20

    So I think I am finally comfortable with the 2.5T Mitsubishi unit as far as the heating load is concerned. Either unit has plenty of cooling sensible capacity, but the 2.5T falls short of meeting our 94F peak latent load of 3,677:

    3T UNIT = 0.75 SHR = 4162 btu/HR latent cooling @ 94F Peak Load
    2.5T UNIT = 0.85 SHR = 2601 btu/HR latent cooling @ 94F Peak Load

    I know the contractor I'm using speaks almost incessantly about removing humidity and latent load capacity, so I'm guessing this will become a talking point if I request the 2.5T system. Any opinions on whether this latent load shortfall is a concern?

  8. Richard_L | | #21

    I'd be a little concerned about how little the 3T unit modulates. A minimum output of 17.4K BTUs at 47º and 9.5K BTUs at 5º is pretty high. If your heat loss is 34K BTUs at 0º, then you're likely going to be short cycling in anything over ~40º. So the 2.5T might be a little better in that respect.

    Though I guess that's kind of expected with this style of heat pump as opposed to slim duct or wall units

    1. dtkrie | | #22

      I'm not certain where you're seeing the 17.4k min. output @ 47F. From the extended performance data, the 3T output @ 47F is 13,800 btu/HR min. I had a manual J ran at 37F; our peak heating load @ 37F is ~16,000 btu, so either the 3T or 2.5T should still be able to run continuous until we start getting temps into the upper 40's. The internal & solar gains generally allow me to turn our furnace off once daytime highs reach into the mid-50's, so I'm not concerned about anything above 50F.

      My current concern is actually the cooling load. One of my old professors, whom I respect greatly and also happened to serve as the president of ASHRAE back around 2015, provided input on our loads and thought the cooling load seemed way too low. I sent him some additional info on our house construction hoping that he perhaps glossed over the finer details.

      1. Expert Member
        Akos | | #23

        SHR is purely a function of CFM/Ton. On the Mitsubishi units you can adjust the external static pressure setting in the thermostat that can reduce the blower speed and lower flow rate. You want to be around 300cfm/ton, so in case of the 2.5T unit with a min cooling of ~14000BTU, you want ~350CFM on low fan.

        Since it is a modulating unit, it has a coil temperature sensor and even if you crank the blower speed way down, it won't frost up the coil. Once you are out of the cooling season you can set the static pressure back to normal.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |